bab-init tries to datalad get a subject from input dataset that's NOT on subject list

[audreycluo]

Summary

This bug is related to issue #176.
When babs-init is run, it does a test to datalad get an example subject from each BIDS dataset. My subject lists are usually a subset of all the subjects in the dataset, so I tried to datalad get just my subjects of interest to save space and time. However, babs-init will try to download a random test subject from the input BIDS dataset and will give an error if it's not there (which it won't be if the subject isn't on my list). It'd be great if babs-init tries to download a subject from the subject list instead of a random one!

Additional details

• BABS version (pip show babs): 0.0.8
• DataLad version (datalad --version): 0.19.6
• Git version (git --version): 2.40.1
• git-annex version (git-annex version; pasting the first printed line is sufficient): 10.20230408-g5b1e8ba77
• datalad-container version (datalad containers-add --version): 1.2.5
• Singularity version (singularity --version): 1.1.8-1.el7

[tientong98]

Hi @audreycluo, it sounds like you have used --list_sub_file in babs-init and that doesn't work?

In the /dir/to/babs/project/analysis/code, there should be a sub_ses_final_inclu.csv (if project is multi-session) or sub_final_inclu.csv (if project is single-session). Does this file only include the people from the list that you want, and not the whole sample?

[audreycluo]

Hi Tien! Thanks so much for looking into this. Yes, I use --list-sub_file. The sub_final_inclu.csv only includes subjects I want, not the whole sample. However, babs-init will try to download a random test subject from the input BIDS dataset even if it's not in sub_final_inclu.csv.